Artificial intelligence (AI) has existed for a while, but it seems it fully shot into the mainstream this year. With this hype has come huge surges in AI-related tech stocks, and one of the main beneficiaries of this has been C3.ai, which is up over 160% for the year through Nov. 24.
C3.ai creates AI applications for enterprises in industries like manufacturing, financial services, defense and intelligence, healthcare, and a handful more. Its main goal is to help organizations operate more efficiently, which, to a certain extent, is the main goal of AI in general.
Although C3.ai is poised to be a major player in the AI space, a better option for investors may be Taiwan Semiconductor Manufacturing Company (TSM -1.75%) because of its unique role in making sure the AI ecosystem functions on a foundational level.
So, what exactly does TSMC do?
TSMC is the world's leading semiconductor (chip) foundry, with over a 56% market share. It essentially pioneered the chip-foundry method where instead of mass producing chips for general sale, it produces chips specifically for each of its clients' needs.
If you're a company building a technology product, you can't purchase TSMC chips straight off the market like you can a company's laptops. Instead, you'd get under contract with TSMC, and then they'll produce the chips based on your specific needs. Apple uses them for the iPhone, Tesla uses them for autonomous driving technologies, and Nvidia and Advanced Micro Devices use them for their graphics processing units (GPUs).
The latter make TSMC's role so important to the AI ecosystem.
The importance of TSMC's chips to AI
AI, for the most part, is built on the premise of having way more knowledge than a single human can have and being able to complete tasks in fractions of the time it would take a human to complete them. For this to work well, vast amounts of data are needed.
Data is to AI what fuel is to a vehicle. AI applications are trained using billions of data points. This huge amount of data isn't just floating in the cloud like your photos and Word documents; it must be kept in data centers specifically built to store and process all this data efficiently. One of the most important pieces to running these data centers are GPUs, which help with the processing power.
No TSMC chips, no GPUs; no GPUs, no data centers; no data centers, no way to train and develop AI applications. So, while TSMC may not directly create AI applications, it's an essential part of the value chain. Of course, TSMC isn't the only chip maker in the world, so the ecosystem would exist without it but would likely not be as efficient.
AI is still a small portion of TSMC's revenue
Despite the boost TSMC's financials could receive from AI-related demand, it hasn't hit that point yet. Here's how its third-quarter 2023 revenue was broken down by segment:
- High-powered computing: 42%
- Smartphone: 39%
- Internet of Things: 9%
- Automotive: 5%
- DCE: 2%
- Others: 3%
During TSMC's Q3 earnings call, it was noted that AI would be about 6% of its revenue this year. However, when asked if AI-related demand would be a big growth driver in the next one to two years, CEO C.C. Wei responded, "Well, the answer is also very simple, yes."
How much of a boost (and its sustainability) has yet to be seen, but TSMC is certainly in a position to benefit from surging demand. It's reasonable to believe GPU makers like Nvidia and AMD are already ramping up their orders with TSMC.
TSMC is fairly priced
Although TSMC's stock has increased over 32% this year, it's still fairly valued, in my opinion. Its price-to-earnings (P/E) ratio is 17.6, which is noticeably less than its five-year average.
TSM PE Ratio data by YCharts.
Slow demand for electronics like smartphones and laptops has weighed on TSMC's financials recently (Q3 revenue was down 14.6% year over year), but that is more of a cyclical issue than an indication of TSMC's business or market position. I think current prices give investors a lot of upside when things turn around.
